Colorado ADMT rules: hearing and last day for written comments · Oct 26, 2026 · See what changes
The process by which a high-risk AI system is checked against the requirements of the EU AI Act before it is placed on the market. For most Annex III systems, providers use internal control (a self-assessment under Annex VI). For biometric systems in point 1 of Annex III, a notified body is involved unless the provider applied harmonised standards or common specifications, and even then the provider may choose one. AI that is a safety component of a product under Annex I follows that product's own conformity procedure. The assessment covers risk management, data governance, technical documentation, logging, transparency, human oversight, accuracy, robustness, and cybersecurity.
Why this matters for your team
For most high-risk categories under the EU AI Act, self-certification is allowed — which means the compliance burden falls on your team, not an external body. Start documenting your risk management and testing processes now so that the self-certification process is a documentation exercise, not a compliance scramble.